A new generation of climate-change experiments: events, not trends

Intensification of weather extremes is currently emerging as one of the most important facets of climate change. Research on extreme events (“event-focused” in contrast to “trend-focused”) has increased in recent years and, in 2004, accounted for one-fifth of the experimental climate-change studies published. Numerous examples, ranging from microbiology and soil science to biogeography, demonstrate how extreme weather events can accelerate shifts in species composition and distribution, thereby facilitating changes in ecosystem functioning. However, assessing the importance of extreme events for ecological processes poses a major challenge because of the very nature of such events: their effects are out of proportion to their short duration. We propose that extreme events can be characterized by statistical extremity, timing, and abruptness relative to the life cycles of the organisms affected. To test system response to changing magnitude and frequency of weather events, controlled experiments are useful tools. These experiments provide essential insights for science and for societies that must develop coping strategies for such events. Here, we discuss future research needs for climate-change experiments in ecology. For illustration, we describe an experimental plan showing how to meet the challenge posed by changes in the frequency or magnitude of extreme events.